We studied the possibility of using artificial intelligence (AI) passive microwave radiometry (MWR) for the diagnostics of venous diseases. MWR measures non-invasive microwave emission (internal temperature) from human body 4 cm deep. The method has been used for early diagnostics in cancer, back pain, brain, COVID-19 pneumonia, and other diseases. In this paper, an AI model based on MWR data is proposed. The model was used to predict the disease state of phlebology patients. We have used MWR and infrared (skin temperature) data of the lower extremities to design a feature space and construct a classification algorithm. Our method has a sensitivity above 0.8 and a specificity above 0.7. At the same time, our method provides an advisory outcome in terms which are understandable for clinicians.
Copyright © 2021 Elsevier B.V. All rights reserved.